[xj-s] Marelli question

XJ-S
181

In reply to a message from Gene Holtzclaw sent Sat 14 Sep 2013:

Gene,

I would think that the 2 twisted wires as a single conductor and
using the outer shielding as the second conductor is still
problematic as the distance between the twisted wires and the outer
shielding varies greatly and would not have a fixed ohm value
consistency as would a 50 ohm coax.

The '89 schematic shows the signal wiring between the power amps
and the PCMI (ECU) to be regular wiring and not coax or twisted
pair - just plain two wires. It is true that the later versions
did upgrade to coax between the PCMI and the power amps. As far as
I can see, the signal from the PCMI to the Lucas EFI ECU has always
been coax with a single conductor and the outer shielding is
grounded.–
lockheed 92 XJS Cpe/97 LT1 Miami FL/ 96 XJS Cv 4.0 Austin TX
Austin, TX, United States
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php –

// please trim quoted text to context only

182

In reply to a message from lockheed sent Sat 14 Sep 2013:

I noticed all of that. The shop that I bought the car from,
that also gave up on this car, offered me the name of who
built this harness. It wasn’t anyone or any company I had
ever heard of. If I have to rebuild it, I actually want to
rebuild it in place, as this harness was built with some
sort of GoreTex looking outer braid, as if they were wanting
to protect it from future fire. It is too long, bulky, and
even more unsightly than the crowd of wiring these cars
originally came with. Maybe I can make it look a little less
bulky looking?–
89 XJS convertible Marelli ignition
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php –

// please trim quoted text to context only

183

As far as I can see, the
signal from the PCMI to the Lucas EFI ECU has always been coax with a
single conductor and the outer shielding is grounded.

What you’re saying is that one conductor has always been grounded?
In that case, if one were using 2-conductor shielded wire, you’d need
to ground the shield AND one of the conductors?

I still don’t believe the coax has anything to do with the problems
here. If it did, neither bank would fire.

– Kirbert

// please trim quoted text to context onlyOn 14 Sep 2013 at 17:56, lockheed wrote:

184

In reply to a message from Kirbert sent Sat 14 Sep 2013:

According to the 1989 wiring schematic, the outer braid on
the wire going to the Lucas unit is grounded. But none of
the others.–
89 XJS convertible Marelli ignition
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php –

// please trim quoted text to context only

185

In reply to a message from Gene Holtzclaw sent Sat 14 Sep 2013:

Gene,

I would expect that the signal voltage from the crank and flywheel
sensors would be relatively small � certainly less than 10-12v, and
probably somewhere between 1-5 volts or less. It would be safe to
set your �0� scope to 10 volts or so and work your way down from
there from a waveform of small amplitude to a larger amplitude by
reducing the voltage scale on your scope from 10 down to 5, and
then 5 down to 1 until you get the amplitude you want for the
waveform, and then read the scale to see exactly what voltage you
have for the waveform. For the power amplifiers to the PCMI (ecu),
start at maybe 15 volts and then reduce the scale range down to
smaller voltages until you get a waveform that is tall enough in
amplitude to resemble what is shown in the JDHT excerpt, and you
are able once again to read a voltage value from the scope trace.

With regard to the frequency of the pulse, remember that the crank
sensor produces 3 pulses per revolution, so at a cranking speed of
300 rpm, the frequency would be 3x300, or 900 Hz. If the idle rpm
is 750 rpm, then the frequency would be 3x750, or 2250 Hz (2.25
Khz).

Since the flywheel has 130 teeth, there would be 130 pulses per
flywheel revolution, so at a cranking speed of 300 rpm, the
frequency would be 300x160, or 48,000 Hz (48 Khz). At a 750 rpm
idle speed, it would be 750x160, or 120,000 Hz (120 Khz). So, in
each case, set the frequency of your scope above the frequency that
you expect to see, and then reduce the frequency scale in steps
until you get width of the pulse trace where you want it.

Now, one can begin to see the importance of using the proper medium
(coax) to connect the sensors with the high frequencies that are
involved � especially at the flywheel. Because the flywheel
frequencies are much higher than the crank frequencies, the
flywheel speed signal will be lost very quickly with improper
wiring.

At 6000 crank rpm: crank signal pulses equal 18,000 Hz (18 Khz),
and the flywheel signal pulses are 960,000Hz (960 Khz). The
flywheel speed signal is in the AM radio band (540 Khz to 1600 Khz)
above 3375 rpm.–
lockheed 92 XJS Cpe/97 LT1 Miami FL/ 96 XJS Cv 4.0 Austin TX
Austin, TX, United States
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php –

// please trim quoted text to context only

186

In reply to a message from lockheed sent Sun 15 Sep 2013:

Thanks Lockheed! It is beginning to make some sense. What I
discovered today, was that I still have the problem. I
switched the front sensor, and did get a signal that was
relative to the rear sensor, but at a much lessor interval.
My problem seemed to change in that the firing of B coil
seemed to last some longer, but still goes away after a few
seconds. I then did what Kirbert had asked about. I
disconnected one of the two wires that were twisted together
in the shielded wire coming from the speed sensors, going to
the ECU, and used the shield as a conductor. I also cut the
grounding loose from those wires. Same problem. I can now
only presume that this car had one sensor that was over the
clearance, (rear), and one sensor that was weak, (front). In
the process of changing out parts, maybe the problem was
made worse by changing out the harness. What I am down to
now, is simply changing out the shielded wire for coaxial,
as I simply am out of ideas on anything else to try. Maybe,
(and I hate to use that term, as I want to KNOW what is
happening) the signal is the same at the sensor connector,
as it is at the ECU, but when the engine turns over a few
revolutions, the ECU ‘‘sees’’ a signal range that is out of
the limits of what the ECU was programmed to see, and it
quits sending a signal. Or a least whatever signal that
signals B bank.–
89 XJS convertible Marelli ignition
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php –

// please trim quoted text to context only

187

In reply to a message from Gene Holtzclaw sent Sun 15 Sep 2013:

My apologies to the group re. Hz which is cycles per second rather
than cycles per minute. Therefore, divided the values I gave for
frequency of the crank and flywheel by 60. Sorry, it was too early
in the morning. Still, the frequencies of the signals are still
quite high and the use of wiring schemes other than coax can still
be a factor.–
lockheed 92 XJS Cpe/97 LT1 Miami FL/ 96 XJS Cv 4.0 Austin TX
Austin, TX, United States
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php –

// please trim quoted text to context only

188

OK, since you’re not actually trying to get the car started just yet,
I have a suggestion for another diagnostic test. Can you just unplug
the connectors from the ignition amps and swap them without swapping
anything else? If the system was in working order that’d mess things
up because the A bank would be firing when the B bank should be
firing, but for diagnostic purposes what I wanna know is whether the
problem switches banks or stays put. If it stays put, the problem is
from the amps on: Amps, coils, or associated wiring. If it swaps
banks, the problem is upstream of there: ECU or associated wiring.

– Kirbert

// please trim quoted text to context only

189

In reply to a message from Kirbert sent Sun 15 Sep 2013:

The good thing about this harness, is that whoever built it,
clearly labeled every connector. A amp, B Amp, A coil, etc.

Switching stuff around was my first wave of attack. No
matter what I do, the problem is always in the B side of the
ignition. And it drops out at the output of the ECU on the B
side. This was found in using the analog volt meter. Now,
using the oscilloscope, found that I had a weak TDC sensor.
I say weak, as it did have a signal, but very faint. The
used one that Mike sent me, sends a signal very much like
the flywheel sensor, although not exactly like it. I do want
to figure out how to get a signal like the one shown in the
factory manual, as this would make me absolutely sure I am
on track here. I just hate to go buy two new sensors (just
yet). I am certain now, that something is happening, that
causes the ECU to stop sending a signal to the B side amp.
If this is due to coax, or shielded wire, causing the ECU to
get a erratic signal input, and causing a repetitive signal
output result, I do not know for sure. What I did do, right
or wrong, is read up on coaxial transmission of signal
verses shielded cable transmission, and they certainly are
two totally different things. The only true coax I have on
this car is the wire going to the Lucas CPU, and it is a
much tighter design than the shielded cables on the harness
I have. It appears that the coax protects the integrity of
the signal from point A to point B, and the shielded wire
protects the signal from outside interference. In using
shielded wire, I’m not sure I am really going to find
anything out without changing the wiring, as I feel as
though I am chasing a ghost here. Paul and I discussed
today, that regardless, it would appear that I will have to
replace the wiring, because even if my coil started firing
today after doing what Kirbert had asked about, its possible
that the car would have not run properly, as the signal
(distortion, integrity?) could change due to improper
transmission of signal once the speed of that signal went up.
I really appreciate all of the input here, and am sure we
all are going to know more than we did due to this,
especially as it comes to the Marelli cars.–
89 XJS convertible Marelli ignition
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php –

// please trim quoted text to context only

190

In reply to a message from Gene Holtzclaw sent Sat 14 Sep 2013:

I just thought I would post this… I found a interesting
article by Phillip Lochner, where he did a dual EDIS 6
system on a 80 XJS using a Megasquirt controller. Evidently,
he too had problems with signal. Here is his post: ‘‘At long
last I’ve ‘‘cured’’ the noisy CLT, IAT and TPS signals. All it
took was NOT earthing these on the engine block but
connecting them to an earth wire coming directly from MS
(megasquirt).’’

I put this in the mix, as this harness has grounds on
literally everything. The original harness did not have
anywhere NEAR the grounds this harness does.–
89 XJS convertible Marelli ignition
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php –

// please trim quoted text to context only

191

Hi All,

In reply to a message from Gene Holtzclaw sent Sat 14 Sep 2013:

I just thought I would post this… I found a interesting article
by Phillip Lochner, where he did a dual EDIS 6 system on a 80 XJS
using a Megasquirt controller. Evidently, he too had problems
with signal. Here is his post:

‘‘At long last I’ve ‘‘cured’’ the noisy CLT, IAT and TPS signals.
All it took was NOT earthing these on the engine block but
connecting them to an earth wire coming directly from MS
(megasquirt).’’

Interesting. What led me to believe that it is a signal problem
solved with true coax cable is when I had to refurbish the connector
doing the interface from the front sensor to where it met up with a
coax cable on top of the engine. That coax cable was not grounded at
all but had a black wire from the sensor/cable interface, on the
coax outer conductor and then went from there on the outside of the
outer conductor, parallel to the coax, I believe all the way back to
the ECU. Can’t say for sure because I did not trace that black
grounding wire back because I didn’t have to. But being an EE and
having worked from 1 mhz all the way up to 94 ghz, my experience told
me that such a ground set up has to do with something about the
signal that was being transmitted all the way back to the ECU.
Depending on the situation it was used many times to get integrity
and fidelity of signal transmission.

–
89 XJS convertible Marelli ignition

Ptipon
Sonora/CA, 90 XJS-V12 conv, United States

// please trim quoted text to context onlyOn Sep 15, 2013, at 5:06 PM, Gene Holtzclaw wrote:

192

In reply to a message from Gene Holtzclaw sent Sun 15 Sep 2013:

To straighten out my previous errors concerning Hz, please accept
the following corrections:

With regard to the frequency of the pulses, remember that the crank

sensor produces 3 pulses per revolution, so at a cranking speed of

300 rpm, the frequency would be 3x300, or 900 cycles/min. If the
idle rpm, is 750 rpm, then the frequency would be 3x750, or 2250
cycles/min - Hz would be 2250/60 = 37.5 Hz.

Since the flywheel has 160 teeth, there would be 160 pulses per

flywheel revolution, so at a cranking speed of 300 rpm, the

frequency would be 300x160, or 48,000 cycles/min - KHz would be
48000/60 = 0.8 Khz.

At a 750 rpm idle speed, it would be 750x160, or 120,000
cycles/min - KHz would be 120,000/60 = 2.0 Khz. So, in

each case, set the frequency of your scope above the frequency that

you expect to see, and then reduce the frequency scale in steps

until you get width of the pulse trace where you want it.

At 6000 rpm, the flywheel pulse frequency would be (6000*160)/60 =
16.0 KHz.

So it is possible that the flywheel speed signal, being of much
higher frequency than the crank signal would be more likely to
degrade much more quickly than the crank signal with improper
wiring. This could explain why the PCMI is unable to compute the B
spark once the rpm picks up and the speed signal degrades enough
that the PCMI cannot compute the B spark.–
lockheed 92 XJS Cpe/97 LT1 Miami FL/ 96 XJS Cv 4.0 Austin TX
Austin, TX, United States
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php –

// please trim quoted text to context only

193

Switching stuff around was my first wave of attack. No
matter what I do, the problem is always in the B side of the
ignition. And it drops out at the output of the ECU on the B
side.

Here’s what I asked: If you connect the B output from the ECU to the
A amp and the A output to the B amp and don’t change anything else,
is it still the B coil that stops firing? Or is it then the A coil?
The two sentences above indicate two different answers; I wanna know
which answer holds up.

I am certain now, that something is happening, that
causes the ECU to stop sending a signal to the B side amp.

I still don’t see what the crank or flywheel sensors could have to do
with that. If they quit working, both the A and B banks would quit
firing.

– Kirbert

// please trim quoted text to context onlyOn 15 Sep 2013 at 16:01, Gene Holtzclaw wrote:

194

In reply to a message from Kirbert sent Mon 16 Sep 2013:

Kirbert, the coil receives its firing signal from the amp.
Switching amps, does move the problem, because the signal
goes from the ECU, into the amp, and then back out, and to
the coil. B signal comes from pins 9 and 10, so no matter
which amp is plugged to it, that is the amp that stops
receiving the signal, and hence, firing stop. I wish the
idea of switching amps helped, because I could easily
replace a amp. It would seem my problem has boiled down to a
input problem. The data the ECU is ‘‘seeing’’ doesn’t fit in
the parameters of its tables(?) and drops one bank off. I
want to feel that its something to do with the front sensor,
since it is the one that controls timing. The rear sensor
controls speed of firing (so it seems). I guess at this
point, for one last stab, is switch my sensors, and see what
happens. The sensor Mike sent me is used, but it could be
borderline bad I guess.

Having never owned a Marelli car, how often to the sensors
go bad?–
89 XJS convertible Marelli ignition
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php –

// please trim quoted text to context only

195

In reply to a message from Gene Holtzclaw sent Mon 16 Sep 2013:

Gene,

I am curious, have you tried scoping the ECU output to the B amp
right at the ECU connector across pins 9 and 10?

By scoping the B output directly at the ECU, it will once and for
all eliminate any possibilty of the harness downstream of the ECU.

As far as upstream (inputs), one thing we have not looked into is
the power supply. If you put a meter across pins 12 and 13 at the
ECU, does the voltage remain constant before and after the B spark
drops off?

I apologize that I am at a loss, not having the JDHT info for the
input pulses you are mapping. But, off hand, the shape of the
front triggers are shaped far different from the flywheel teeth
providing the rear triggers…so I don’t think they will ever look
the same on the scope, although I would think the amplitudes could
be identical. If the amplitudes are different, I would also think
you can adjust that by adjusting the air gaps.

Finally, the sensors on my car have never gone bad.

John–
John. '95 XJS 6.0L convertible. Southlake, TX
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php –

// please trim quoted text to context only

196

In reply to a message from Gene Holtzclaw sent Mon 16 Sep 2013:

‘‘The rear sensor controls speed of firing (so it seems).’’

Gene, according to the documentation from AJ6 Engineering and JDHT,
the flywheel sensor merely provides an engine speed reference (rpm)
to the PCMI. The speed and timing reference of the firing signals
are provided by the front timing fingers passing the front sensor.–
lockheed 92 XJS Cpe/97 LT1 Miami FL/ 96 XJS Cv 4.0 Austin TX
Austin, TX, United States
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php –

// please trim quoted text to context only

197

In reply to a message from CJ95 sent Mon 16 Sep 2013:

''By scoping the B output directly at the ECU, it will once and for

all eliminate any possibilty of the harness downstream of the ECU.’’

Yes, but it could also indict the harness if you find that the
signal at the power amp is degraded from the signal at the PCMI.

‘’ But, off hand, the shape of the front triggers are shaped far
different from the flywheel teeth providing the rear triggers…’’

No, they are not ‘‘far’’ different, but there is a slight difference
in the waveform, but the biggest difference is of frequency per
crank shaft revolution - 3 pulses per revolution vs 160 pulses per
revolution. See the waveforms for each from JDHT that I have sent
you off topic.–
lockheed 92 XJS Cpe/97 LT1 Miami FL/ 96 XJS Cv 4.0 Austin TX
Austin, TX, United States
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php –

// please trim quoted text to context only

198

I am curious, have you tried scoping the ECU output to the B amp right
at the ECU connector across pins 9 and 10?

I dunno the inner workings of this thing, but I wouldn’t expect the
signal to be between 9 and 10. I’d expect the signal to be between 9
and ground, or POSSIBLY between 9 and +12V (Jaguar likes ground
switching stuff!).

By scoping the B output directly at the ECU, it will once and for all
eliminate any possibilty of the harness downstream of the ECU.

Actually, maybe not. If pin 10 is actually a feedback signal from
the amp back to the ECU, then a fault in that circuit could be
confusing the ECU.

As far as upstream (inputs), one thing we have not looked into is the
power supply. If you put a meter across pins 12 and 13 at the ECU,
does the voltage remain constant before and after the B spark drops
off?

I asked that early on. I certainly HOPE that power supply has been
verified by now!

Perhaps a better avenue for investigation would be to scope the
difference between the #4 terminal on each amp and pin 13 on the ECU,
as well as between #2 on each amp and pin 12 on the ECU, just to
check for stray voltages. Both those checks should come up with
absolutely nothing, but if there’s noise it’ll probably cause
trouble.

But, off hand, the shape of the front
triggers are shaped far different from the flywheel teeth providing
the rear triggers…

Are these the kind of things where a ferrous piece of metal is coming
closer and then is moving farther away, and it’s the point of
switching polarity – from getting closer to moving farther away –
that provides the actual triggering?

– Kirbert

// please trim quoted text to context onlyOn 16 Sep 2013 at 7:09, CJ95 wrote:

199

In reply to a message from lockheed sent Mon 16 Sep 2013:

I was thinking we already know that the signal is degraded
at the amp, or the B spark would continue sparking.
Checking at the ECU defines the problem as an ECU input,
vs. a harness problem in the output.

Alan sent me the waveforms. Their shape explains why the
shield wire is not grounded.–
The original message included these comments:

Yes, but it could also indict the harness if you find that the
signal at the power amp is degraded from the signal at the PCMI.

–
John. '95 XJS 6.0L convertible. Southlake, TX
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php –

// please trim quoted text to context only

200

In reply to a message from CJ95 sent Mon 16 Sep 2013:

‘‘I was thinking we already know that the signal is degraded at the
amp, or the B spark would continue sparking. Checking at the ECU
defines the problem as an ECU input, vs. a harness problem in the
output.’’

Just to make sure we are on the same page here, we suspect that the
speed signal output from the flywheel sensor arriving at pins 3 and
16 of the PCMI has been degraded by improper wiring in the
harness. Likewise, we can suspect that the crank sensor signal to
the PCMI (pins 1 and 2) may be degraded for the same reason, but to
a lesser degree because of the lower frequency. Bad stuff in, bad
stuff out. Only when we are sure that what is going into the PCMI
is good, can we expect proper output. Mike has already verified
that your PCMI is working correctly.–
lockheed 92 XJS Cpe/97 LT1 Miami FL/ 96 XJS Cv 4.0 Austin TX
Austin, TX, United States
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php –

// please trim quoted text to context only